We propose to develop techniques to quantify trabecular bone density and structure using non-invasive Magnetic Resonance (MR) techniques. When a sample consisting of trabecular bone and marrow is placed in a static magnetic field, the difference between the magnetic susceptibility of trabecular bone and bone marrow gives rise to distortions in the magnetic lines of force within the sample. In magnetic resonance these distortions in the lines of force generate magnetic field inhomogeneities. The behavior of marrow fat and water protons (1H) is modified by these magnetic field inhomogeneities and the net transverse magnetization in gradient echo magnetic resonance images decays faster, resulting in a decrease in the marrow transverse relaxation time, T2*, and consequently a reduction in signal intensity of bone marrow. The extent of T2* reduction and loss of signal intensity is directly proportional to the concentration of discontinuities in magnetic susceptibility, i.e. trabecular bone density, and also depends on the spatial distribution of trabecular bone and the imaging field strength. In this study we will quantify the reduction in relaxation time T2* of marrow or marrow equivalent material present in the intra-trabecular spaces using dried human specimens and fresh specimens from cadavers. We will relate the rate of change of relaxation time with the density of the surrounding trabecular bone measured using quantitative computed tomography. The relative role of bone marrow composition in the variation of T2* with bone density will be assessed and the MR estimated parameters will be compared to the mechanical strength of the specimens. Image analysis techniques will be developed in order to characterize the trabecular pattern. The perimeter-to-area ratio of the trabecular bone in two dimensions, trabecular spacing and the texture of the trabecular pattern assessed using a measure of the fractal dimension will be derived from these MR images. These parameters will be evaluated along different planes, and trabecular bone anisotropy will be deduced. The correlation between these measures of trabecular structure and mechanical strength of the specimens and histomorphometric estimates of the mean trabecular spacing and dimensions will be derived. After establishing the effectiveness of these techniques in assessing trabecular bone density, structure and its' correlation with histomorphometry and mechanical strength, we will conduct a case control study to determine the efficacy of these outcome variables as predictors of fracture risk. Postmenopausal women (60-70 years) with and without vertebral fractures will be recruited for the purpose of this study. Trabecular bone density of the lumbar spine will be assessed using quantitative computed tomography and the relaxation time T2* distribution will be measured using magnetic resonance imaging techniques at 1.5 Tesla, along with the image analysis techniques developed. Logistic regression will be used to determine whether these outcome variables are better predictors of relative fracture risk compared to measures of bone density alone.